Bridge Safety State Classification Based on Unsupervised Machine Learning

With the development of artificial intelligence and big data technology, it is becoming a research trend to combine structural health monitoring (SHM) with these new techniques in order to make the analysis more intelligent. Due to the complexity and uncertainty of the structure environments, there are still problems such as low feature extraction efficiency and low accuracy in the research on the multivariable correlation features for structural perception data. Recently, the advantages of generative adversarial networks for nonlinear feature extraction have been noticed and this method has the potential to be applied to structural health monitoring. In this paper, an abnormal state detection model of monitoring data will be designed through the generative adversarial network of unsupervised learning. The abnormal state caused by bridge damage will be detected from a data-driven perspective, which provides support for the classification of bridge safety state abnormalities.